Distance dependence of photoinduced electron transfer in metalloporphyrin dimers

Article abstract of DOI:10.1021/jp991766s

To study the dynamics and mechanism of intramolecular photoinduced electron transfer (PET) reactions, a series of (Zn^II-Fe^III) mejo-tetraarylmetalloporphyrin dimers were synthesized and the kinetics of their PET reactivity was measured. Molecular building blocks were prepared by selective nucleophilic aromatic substitution of a para fluorine on tetraarylporphyrins containing a single pentafluorophenyl group. This synthetic approach allows a wide variety of systematic modifications such as type and length of spacer, metal center, and redox-potential difference between donor and acceptor. The edge-to-edge distance between the two porphyrins varies from 14.4 to 27.3 ?. Into a symmetric dimer, with two identical porphyrins covalently linked by a rigid partly saturated bridge, one zinc(II) and one iron(III) can be inserted. From measurements of fluorescence lifetimes the rate constants for PET from the electronically excited state of the zinc porphyrin to the bis(imidazole)iron porphyrin cation were evaluated. The electron-transfer rate decreases by a factor of only 165 when the distance increases by 13 ?. This small decrease is indicative of a surprisingly weak attenuation of the electronic coupling with distance.

Full text of DOI:10.1021/jp991766s

10540
J. Phys. Chem. A 1999, 103, 10540-10552
Distance Dependence of Photoinduced Electron Transfer in Metalloporphyrin Dimers†
Carmita F. Portela,‡,§ Jarmila Brunckova,‡ Joseph L. Richards,‡,| Bernd Scho1llhorn,‡,
Yassuko Iamamoto,∇ Douglas Magde,‡ Teddy G. Traylor,‡,X and Charles L. Perrin*,‡
Department of Chemistry and Biochemistry, UniVersity of CaliforniasSan Diego, 9500 Gilman DriVe,
La Jolla, California 92093-0506, UniVersidade Federal de Pernambuco, Recife, Brazil,
UniVersidade de Sa˜o Paulo, Ribeira˜o Preto, Brazil
ReceiVed: June 2, 1999; In Final Form: September 7, 1999
To study the dynamics and mechanism of intramolecular photoinduced electron transfer (PET) reactions, a
series of (ZnII-FeIII) meso-tetraarylmetalloporphyrin dimers were synthesized and the kinetics of their PET
reactivity was measured. Molecular building blocks were prepared by selective nucleophilic aromatic
substitution of a para fluorine on tetraarylporphyrins containing a single pentafluorophenyl group. This synthetic
approach allows a wide variety of systematic modifications such as type and length of spacer, metal center,
and redox-potential difference between donor and acceptor. The edge-to-edge distance between the two
porphyrins varies from 14.4 to 27.3 Å. Into a symmetric dimer, with two identical porphyrins covalently
linked by a rigid partly saturated bridge, one zinc(II) and one iron(III) can be inserted. From measurements
of fluorescence lifetimes the rate constants for PET from the electronically excited state of the zinc porphyrin
to the bis(imidazole)iron porphyrin cation were evaluated. The electron-transfer rate decreases by a factor of
only 165 when the distance increases by 13 Å. This small decrease is indicative of a surprisingly weak
attenuation of the electronic coupling with distance.
Introduction
ET occurs through the bonds of the bridge and not through the
surrounding medium. Often ET appears to be mediated by a
superexchange mechanism involving the antibonding orbitals
of the bridge.18 Therefore, aromatic and unsaturated bridges are
expected to be more effective than fully saturated bridges.
A recent study of three supramolecular porphyrin systems
allowed a comparison of the electronic coupling provided by
σ, π, and hydrogen bonds.19 Contrary to general expectations,
the electronic coupling across a hydrogen-bond interface is
greater than that across an analogous interface composed entirely
of carbon-carbon σ bonds. Moreover, only a two-fold increase
in the ET rate was observed in a porphyrin dimer with a
π-containing cis-bicyclo[3.3.0]octa-2,6-dien-3,7-ylidene (BCOE),
as compared to the saturated cis-bicyclo[3.3.0]octan-3,7-ylidene
(BCO) linker.
To test how the efficiency of ET is related to the molecular
structure of the spacer, systems with rigid, nonaromatic bridges
that impose a well-defined molecular geometry should be
investigated. Through the precise experimental control of
geometry, we hope to provide further experimental information
for refinement of theoretical treatments. We now report the
results of photoinduced ET in a series of metalloporphyrin
dimers rigidly connected by systems including aliphatic rings.
The saturated rings used here offer the advantages of high
chemical stability and superior insulating characteristics, which
we expect will minimize direct interactions between the donor
and acceptor chromophores. The minimal interaction is impor-
tant for analogy to photosynthetic reaction centers, where the
key chromophores are separated by large distances.
Over the past decade substantial theoretical and experimental
effort has been directed toward understanding the dynamics of
electron transfer (ET) in both proteins and model compounds.1
Remarkable progress has been achieved through experiments
with defined distances, energetics, and orientations of electron
donor and acceptor in supramolecules. Nevertheless, an incom-
pletely solved problem is the delineation of the factors that
control the dependence of ET rates on the distance between
the electron donor and the electron acceptor.2 A theoretical
model has been developed for electron tunneling across
proteins.3 The distance dependence of ET across DNA is
currently a matter of considerable controversy.4
Covalently linked porphyrin-quinone diads have been ex-
tensively studied from the point of view of photoinduced
electron transfer (PET).5-13 However, considerably less infor-
mation is available for two porphyrins covalently linked by a
rigid spacer.14,15 McLendon and co-workers studied PET in a
series of Zn,Fe protoporphyrin dimers where the bridge consists
of either one, two, or three para-linked phenyl groups.16 They
found a weak distance dependence for the ET rates, which was
interpreted in terms of a moderate loss of conjugation at each
phenyl junction. This result can be described by a simple theory
that assumes delocalization within the phenyl ring.
The molecular structure of the bridge plays an important role
in ET between donor and acceptor moieties.17 In most cases
† Dedicated to Kent R. Wilson.
* To whom correspondence should be addressed. E-mail: cperrin@
ucsd.edu. Fax: (858)-822-0386.
The development of a method20 for the preparation of
functionalized polyhalogenated tetraarylporphyrins by selective
substitution of para fluorines of meso-tetrakis(pentafluorophe-
nyl)porphyrins led us to a series of new supramolecular systems
12a-g illustrated in Figure 1. These include a wide variety of
‡ University of CaliforniasSan Diego.
§ Universidade Federal de Pernambuco.
| Present address: Mesa State College, Grand Junction, CO 81502.
Present Address: Ecole Normale Superieure, 75231 Paris, France.
∇ Universidade de Sa˜o Paulo.
X Deceased, June 23, 1993.
10.1021/jp991766s CCC: $18.00 © 1999 American Chemical Society
Published on Web 11/09/1999

Electron Transfer in Metalloporphyrin Dimers
J. Phys. Chem. A, Vol. 103, No. 49, 1999 10541
from Aldrich, Sigma, or Fluka and used as received unless
otherwise noted. THF was distilled under argon from sodium
benzophenone ketyl immediately prior to use. Pyrrole was
distilled from CaH2 under reduced pressure prior to use. N,N′-
Dimethylpropanediamine (Aldrich) was dried over molecular
sieves. Solvents for physical measurements were of spectro-
photometric grade.
General Methods. 1H, 19F, and 13C NMR spectra were
measured on either a 500-MHz Varian Unity or a 300-MHz
GE QE spectrometer, in CDCl3 with tetramethylsilane (TMS)
(1H, 13C) or hexafluorobenzene (19F) as the internal standard
unless noted otherwise. UV-visible spectra in dichloromethane
were recorded on a Kontron Uvikon-810 spectrophotometer.
Molar absorption coefficients were determined on 2-3 mg of
a compound, weighed accurately, and diluted to an appropriate
absorbance. All reactions were carried out under an argon
atmosphere in purified solvents with shielding from light and
were followed by thin-layer chromatography (silica gel on
aluminum foil, Merck 5554). Preparative column chromatog-
raphy was performed on silica gel (Davisil grade 643 type 150A,
230-435 mesh) or neutral alumina (Brockman Activity I, 60-
325 mesh). Mass spectral analyses were conducted by The
Scripps Research Institute Mass Spectrometry Facility, La Jolla,
CA, by FAB+, EI+, or electrospray+ techniques. These mol-
ecules are so large that the peak calculated for the lightest
isotopes is not the most prominent, so that sometimes only the
most abundant is reported, and this does agree with the
calculated peak. Microanalyses were performed by Desert
Analytical, Tucson, AZ.
Samples for fluorescence measurements were prepared at
concentrations of ca. 1 × 10-5 M in dichloromethane containing
0.01 M 1-methylimidazole. The excitation source was a mode-
locked argon ion laser coupled to a synchronously pumped
cavity-dumped dye laser operating at 568 nm. Fluorescence
decay at 610, 650, or 670 nm was monitored by time-correlated
single-photon counting (TCSPC).21,22 Further details have been
described elsewhere.22
The fluorescence lifetime of the reference dimer is given by
eq 1, where kF, kISC, and kIC are the rate constants for
τ0 ) 1/(kF + kISC + kIC)
(1)
Figure 1. Porphyrin dimers for study of the dependence of ET rate
on distance.
fluorescence, intersystem crossing, and nonradiative internal
conversion, respectively. If the only additional deactivation
pathway available for the Zn,Fe derivatives is the electron
transfer and if the rate constants for all of the other deactivation
pathways are the same as those in the corresponding reference
dimer, the fluorescence lifetimes of the Zn,Fe dimers are given
by eq 2. This assumes that energy transfer does not contribute
spacers, namely, piperazyl (pip), bicyclo[2.2.2]octane-1,4-
diamine (bcoda), 4,4′-bipiperidyl (bip), the steroidal diamine
5R-androstanyl-3â,17â-diamine (stda), the steroidal diether 5R-
androstanyl-3R,17â-diether (stde), bis(piperazyl)perfluorobi-
phenylyl (pip-pfbp-pip), and N,N′-dimethyl-1,3-diaminopropyl
(dmdap), each linked to the two porphyrins by a tetrafluorophe-
nyl group. The further tasks are to insert two different metals,
so as to allow PET, and to characterize the various porphyrin
derivatives, which is not trivial for substances of such high
molecular weight. These porphyrin dimers then permit the study
of the dependence of electron-transfer rates on distance and on
the molecular structure of the intervening linkage. In this series
the other critical parameters governing PET rates, such as
driving force, reorganization energy, and orientation, can be kept
constant. In particular, the electron transfer is always from a
singlet excited-state Zn porphyrin to the six-coordinate low-
spin Fe(III) complex of the identical porphyrin.
τ ) 1/(kF + kISC + kIC + kET)
(2)
to the deactivation,16a,19 an issue that is further addressed below.
It then follows that the electron-transfer rate constant kET can
be obtained from the observed fluorescence lifetimes as in eq
3. Experimental values are averaged over data sets from different
kET ) 1/τ - 1/τ0
(3)
preparations on at least two different days, with at least two
measurements at each of two wavelengths on each day.
Distance Measures. It is necessary to establish a measure
of the spatial distance between the donor and acceptor in these
porphyrin dimers. We have chosen the distance between
porphyrin edges. The through-space distance (dspace) is the length
Experimental Section
Materials. All solvents were purchased from Fisher, dried,
and distilled by standard methods. All reagents were purchased

10542 J. Phys. Chem. A, Vol. 103, No. 49, 1999
Portela et al.
of the vector connecting the meso-carbons of the zinc and iron
porphyrins that are bonded directly to the bridge. The through-
bond distance (dbond) is the sum of the lengths of the bonds
connecting the meso-carbons of the zinc and iron porphyrins
that are bonded directly to the bridge. The number of bonds
present in the shortest pathway across the bridge connecting
the zinc and iron porphyrins at their meso-carbon atoms is the
total bond pathway (Nbonds). The through-space and through-
bond distances were calculated by using a Monte Carlo
conformation search and MM2 force field. For the more
complex stde and stda bridges of 8d, 12d, 8e, and 12e, dspace
corresponds to an average over representative conformers
selected from a series of low-energy minima.
Even though the linkages may permit rotation of the por-
phyrins relative to the bridge, and also within the bipiperidyl
and bis(piperazyl)perfluorobiphenylyl bridges, those rotations
do not greatly change the distance between the two porphyrin
edges. This is a consequence of the rigidity of the piperazine,
bicyclooctane, and piperidine rings. For the bicyclooctane and
steroidal linkages, rotation around the inner C-N or C-O bonds
could lead to conformational heterogeneity, if it is slow relative
to fluorescence, but the data below show that this complication
does not intrude.
5-(Pentafluorophenyl)-10,15,20-triphenylporphinatozinc
(ZnTPPF5, 2). A solution of 1 (0.43 g, 0.61 mmol) and zinc
acetate dihydrate (0.81 g, 3.68 mmol) in dichloromethane (90
mL) plus methanol (90 mL) was heated to reflux for 5 h under
argon in the dark. The reaction mixture was then cooled to room
temperature and concentrated approximately three-fold by
evaporation of solvent in vacuo. The residue was diluted with
dichloromethane (600 mL) and washed with water (3 × 100
mL), saturated sodium bicarbonate solution (3 × 100 mL), and
brine (3 × 100 mL). The water layers were combined and
rewashed with dichloromethane (2 × 50 mL). The combined
organic layer was dried over sodium sulfate, filtered, and
evaporated in vacuo to dryness. The crude product was purified
by column chromatography (silica gel: petroleum ether/dichlo-
romethane ) 4/1) to yield 2 (0.460 g, 97%) as a pinkish-purple
solid: 1H NMR δ 9.02 (d, 2H, J ) 5 Hz, pyrrole), 8.94 (d, 4H,
J ) 3.5 Hz, pyrrole), 8.85 (d, 2H, J ) 5 Hz, pyrrole), 8.35-
8.15 (m, 6H), 7.88-7.70 (m, 9H); 19F NMR δ 24.62 (dd, 2F, J
) 24 and 8 Hz, o), 8.41 (t, 1F, J ) 24 Hz, p), -0.76 (dt, 2F,
J ) 24 and 8 Hz, m); UV λmax 416 (s), 547 (w), 585 (w) nm.
Anal. Calcd for C44H23N4F5Zn: C, 68.84; H, 3.02; N, 7.29.
Found: C, 69.08; H, 2.86; N, 7.37.
5-(Pentafluorophenyl)-10,15,20-tris(2,6-dichlorophenyl)-
21H,23H-porphyrin (H2TPPCl6F5, 3).23 A 3-L three-neck,
round-bottomed flask, equipped with a reflux condenser and
an argon inlet, was charged with dry dichloromethane (1400
mL), 2,6-dichlorobenzaldehyde (2.45 g, 14.0 mmol), pentafluo-
robenzaldehyde (0.25 mL, 2.0 mmol), and pyrrole (1.14 mL,
16.0 mmol). The mixture was stirred and purged with argon
for >20 min, after which BF3‚Et2O (40 mL) was added. The
solution rapidly turned dark yellow and then orange-red. The
mixture was stirred for 12 h at room temperature. Another
portion of BF3‚Et2O (20 mL) was added, and the solution was
stirred for 24 h more. Tetrachloro-1,4-benzoquinone (p-chloranil;
3.93 g, 20 mmol) was then added to the dark solution. After
refluxing for 4 h, the green solution was concentrated in vacuo
to about 100 mL. Neutral alumina (10-20 g) was added, and
the residual solvent was removed by evaporation in vacuo. This
material was first flash chromatographed (neutral alumina:
toluene/dichloromethane ) 1/1), concentrated, and then purified
by gradient column chromatography (neutral alumina: petro-
leum ether/dichloromethane ) 3/1 gradually to 1/1) to yield 3
(314 mg, 17%) as purple crystals after recrystallization from
dichloromethane/methanol: 1H NMR δ 8.77 (m, 4H, pyrrole),
8.70 (m, 4H, pyrrole), 7.83-7.79 (m, 6H, m), 7.74-7.68 (m,
3H, p), -2.61(s, 2H, NH); 19F NMR δ 25.40 (dd, 2F, J ) 24
and 8 Hz, o), 9.25 (t, 1F, J ) 24 Hz, p), -0.44 (dt, 2F, J ) 24
and 8 Hz, m); UV λmax (10-4ꢀ) 416 (25.6), 512 (1.63), 540 (sh),
587 (0.53), 642 (0.25) nm; MS (CI+, NH3) m/z 911 (M+, 100%).
Anal. Calcd for C44H19Cl6F5N4: C, 57.99; H, 2.10; N, 6.15.
Found: C, 57.89; H, 2.12; N, 6.05.
An alternative is to choose the metal-metal distance, although
this is conformation-dependent and less well-defined. Besides,
the donor wave function of 1ZnP* and the acceptor wave
function of FeIIIP[bis(imidazole)] are both delocalized over their
entire porphyrin rings. Therefore, it is reasonable to measure
distances in porphyrin dimers simply between porphyrin edges.
Porphyrin Syntheses and Characterization
5-(Pentafluorophenyl)-10,15,20-triphenyl-21H,23H-por-
phine (H2TPPF5, 1). Freshly distilled dichloromethane (2.5 L)
in a two-neck, round-bottomed flask protected from light and
fitted with septa was stirred and purged with argon for 1 h at
room temperature. Then benzaldehyde (5.0 mL, 49 mmol),
pentafluorobenzaldehyde (0.86 mL, 7.0 mmol), pyrrole (4.0 mL,
59 mmol), and 2,2-dimethoxypropane (4.0 mL, 33 mmol) were
added via syringe in this order. The stirring and bubbling with
argon was continued for an additional 15 min, when boron
trifluoride etherate (1.6 mL, 13 mmol) was added via syringe
in two portions at 30-min intervals. After this addition was
completed, the reaction mixture changed immediately from
colorless to red-brown. The dark reaction mixture was then
stirred for 1.5 h with a slow stream of argon passing through
the solution. A solution of 2,3-dichloro-5,6-dicyanobenzo-
quinone (DDQ; 10.2 g, 45 mmol) in freshly distilled toluene
(100 mL) was then added via a cannula needle. The mixture
was stirred for an additional 5 h. A longer condensation or
oxidation time leads to a larger amount of polymeric material,
decreases the overall yield, and causes purification difficulties.
The solvent was evaporated to dryness in vacuo. The crude
mixture (purple-black powder) was flash chromatographed
(neutral alumina: toluene/dichloromethane ) 1/3) to remove
polar substances. The mixture of products was purified by
gradient column chromatography (neutral alumina: petroleum
ether/dichloromethane ) 6/1 gradually to 3/1) to yield 5,10,-
15,20-tetrakis(pentafluorophenyl)-21H,23H-porphine (0.07 g,
2%) as purple crystals, 5,10,15,20-tetraphenyl-21H,23H-por-
phine (0.16 g, 7.4%), and the desired compound 1 (0.71 g, 29%)
as purple crystals: 1H NMR δ 8.95-8.75 (m, 8H), 8.25-8.20
(m, 6H), 7.81-7.70 (m, 9H), -2.57 (s, 2H, NH); UV λmax
(10-4ꢀ) 396 (sh) (8.61), 416 (42.3), 513 (1.9), 548 (0.57), 588
(0.58), 643 (0.26) nm. Anal. Calcd for C44H25N4F5: C, 74.99;
H, 3.58; N, 7.95. Found: C, 75.29; H, 3.55; N, 7.97.
5-(Pentafluorophenyl)-10,15,20-tris(2,6-dichlorophenyl)-
porphinatozinc (ZnTPPCl6F5, 4). 3 (243 mg, 0.266 mmol) was
dissolved in a mixture of dichloromethane (40 mL) and methanol
(40 mL). Zinc acetate dihydrate (580 mg, 2.66 mmol) was
added, and the reaction mixture was refluxed for 12 h. After
cooling to room temperature, the solution was neutralized with
a saturated aqueous sodium bicarbonate solution. The organic
solvents were evaporated in vacuo. Dichloromethane (200 mL)
was added, and the organic layer was washed with water (2 ×
50 mL) and brine (2 × 50 mL), dried over sodium sulfate,
filtered, and evaporated in vacuo. The residue was recrystallized
from dichloromethane/hexane to provide zinc complex 4 (255
mg, 98%) as red-purple crystals: 1H NMR δ 8.85 (m, 4H,

Electron Transfer in Metalloporphyrin Dimers
J. Phys. Chem. A, Vol. 103, No. 49, 1999 10543
pyrrole), 8.80 (m, 4H, pyrrole), 7.80 (m, 6H, m), 7.74-7.68
(m, 3H, p); 19F NMR δ 25.15 (dd, 2F, J ) 24 and 8 Hz, o),
8.67 (t, 1F, J ) 24 Hz, p), -0.76 (dt, 2F, J ) 24 and 8 Hz, m);
UV λmax (10-4ꢀ) 418 (72.4), 510 (sh, 0.35), 548 (2.93), 582
(sh, 0.39) nm; MS (CI+ NH3) m/z 975 (M+, 100%). Anal. Calcd
for C44H17Cl6F5N4Zn: C, 54.22; H, 1.76; N, 5.75. Found: C,
54.05; H, 1.97; N, 5.70.
and washed with water (3 × 20 mL), a saturated sodium
bicarbonate solution (3 × 20 mL), and brine (3 × 20 mL). The
water layers were combined and washed with dichloromethane
(2 × 40 mL). The combined organic layer was dried over
sodium sulfate, filtered, and evaporated in vacuo, and the
residual 1-methyl-2-pyrrolidinone was removed by vacuum
distillation at 60 °C. Purification by column chromatography
(silica gel: dichloromethane/methanol/ammonium hydroxide )
15/1/0.1) provided 7c (158 mg, 70%) as a dark purple solid:
1H NMR δ 8.98-8.89 (m, 8H), 8.26-8.23 (m, 6H), 7.76-7.74
(m, 9H), 3.48 (d, 2H, J ) 11 Hz), 3.01 (dd, 2H, J ) 11 Hz),
1.40-1.28 (m, 1H), 0.99-0.87 (m, 3H), 0.77-0.66 (m, 2H),
0.14-0.11 (bm, 1H), -0.25 to -0.48 (bm, 2H), -0.49 to -0.68
(bm, 1H), -1.02 to -1.44 (bm, 1H), -2.00 to -2.32 (bm, 2H),
-2.41 to -2.80 (bm, 1H); 13C NMR δ 150.62, 150.27, 149.99,
143.37, 143.33, 134.63, 134.57, 132.87, 132.03, 131.71, 130.11,
127.38, 126.41, 122.21, 121.15, 51.67 (2CH2), 42.01 (2CH2),
39.56 (CH), 38.25 (CH), 29.22 (2CH2), 25.33 (2CH2); 19F NMR
δ 22.48 (bs, 2F), 10.54 (bs, 2F); UV λmax 419 (s), 548 (w), 587
(w) nm. HRMS. Calcd for C54H43N6F4Zn [(M + H)+]: 915.2777.
Found: 915.2747.
5-(Pentafluorophenyl)-10,15,20-triphenylporphinatoiron-
(III) Acetate (FeTPPF5‚OAc, 5). A mixture of 1 (100 mg, 0.14
mmol) and ferrous bromide (42 mg, 0.45 mmol) in 10 mL of
freshly distilled tetrahydrofuran (THF) was heated to reflux
under argon for 2 h. After cooling to room temperature, solvent
was evaporated in vacuo and the residue was dissolved in
dichloromethane (50 mL), washed with 10% aqueous acetic acid
(3 × 15 mL) and brine (3 × 15 mL), and dried over calcium
chloride. The solvent was evaporated in vacuo. The residue was
crystallized from dichloromethane/hexanes to provide 5 (100
mg, 90%) as a brown-purple solid: UV λmax 375 (sh, w), 415
(s), 509 (w), 580 (w), 648 (w) nm. HRMS m/z calcd for
C46H27N4F5FeO2 [(M + H)+]: 818.1404. Found: 818.1401.
5-(Pentafluorophenyl)-10,15,20-tris(2,6-dichlorophenyl)-
porphinatoiron(III) Acetate (FeTPPCl6F5‚OAc, 6). A mixture
of 3 (100 mg, 0.11 mmol) and ferrous bromide (42 mg, 0.45
mmol) in 10 mL of freshly distilled dimethylformamide (DMF)
was heated to reflux under argon for 0.5 h. After cooling to
room temperature, the mixture was diluted with dichloromethane
(50 mL), washed with 10% aqueous acetic acid (3 × 15 mL)
and brine (3 × 15 mL), and dried over sodium sulfate. The
solvent was evaporated in vacuo, and the remaining DMF was
removed by vacuum distillation. The crude mixture was purified
by column chromatography (alumina: dichloromethane/metha-
nol ) 100/1) to give 6 (91 mg, 81%) as a purple-brown solid:
UV λmax (10-4ꢀ) 380 (sh, 5.54), 416 (8.92), 511 (1.20), 582
(0.37), 649 (0.33) nm. MS. Calcd for C46H21N4F5FeO2Cl6 [(M
+ H)+]: 1025. Found: 1025.
ZnTPPF4-stde (7d) and Zn,Zn(TPPF4)2-stde (8d). Com-
mercially available 5R-androstane-3R,17â-diol (8.8 mg, 0.03
mmol) was added to a mixture of sodium hydride (2.7 mg, 0.07
mmol, 60% dispersion in paraffin) and 1-methyl-2-pyrrolidinone
(2 mL) at 5 °C under argon. After 1 h 2 (50 mg, 0.065 mmol)
was added, and the reaction mixture was heated to 50 °C for
48 h. After cooling to room temperature, the mixture was poured
onto ice water (50 mL) and extracted with dichloromethane (3
× 15 mL). The combined organic layer was dried over sodium
sulfate and evaporated to dryness. Purification by column
chromatography (silica gel: petroleum ether/dichloromethane
) 1/1) gave two compounds, 7d and 8d, as purple solids. For
the more polar 7d (6 mg, 19%): 1H NMR δ 9.05-8.90 (m,
8H), 8.25-8.20 (m, 6H), 7.85-7.70 (m, 9H), 4.70-4.60 (m,
1H), 3.30-3.20 (m, 1H), 2.45-2.30 (m, 2H), 2.20-0.52 (m,
26H); UV λmax 418 (s), 548 (w), 584 (w) nm. MS. Calcd for
C63H54N4F4O2Zn: 2082 (2 × MH+). Found: 2082 (2 × MH+).
For the less polar 8d (21 mg, 40%): 1H NMR δ 9.10-8.90
(m, 16H), 8.30-8.20 (m, 12H), 7.85-7.70 (m, 18H), 5.05-
4.95 (m, 1H), 4.75-4.60 (m, 1H), 2.60-0.55 (m, 28H); UV
ZnTPPF4-pip (7a). A solution of zinc porphyrin 2 (100 mg,
0.13 mmol) and piperazine (112 mg, 1.3 mmol) in freshly
distilled dry 1-methyl-2-pyrrolidinone (1.5 mL) was heated to
110-120 °C for 20 h under argon. The reaction mixture was
then cooled to room temperature, diluted with dichloromethane
(50 mL), and washed with water (3 × 20 mL), a saturated
sodium bicarbonate solution (3 × 20 mL), and brine (3 × 20
mL). The water layers were combined and extracted with
dichloromethane (2 × 40 mL). The combined organic layer was
dried over sodium sulfate, filtered, and evaporated in vacuo.
The residual 1-methyl-2-pyrrolidinone was removed by vacuum
distillation at 60 °C. Purification by column chromatography
(silica gel: dichloromethane/methanol/ammonium hydroxide )
λmax 418 (s), 547 (w), 586 (w) nm. MS. Calcd for C107H77N8F8O2-
Zn2 [(M + H)+]: 1785. Found: 1785.
ZnTPPF4-pip-pfbp (7f). A mixture of 7a (20.0 mg, 0.024
mmol), perfluorobiphenyl (200 mg, 0.60 mmol), N,N-diisopro-
pylethylamine (0.2 mL), and dimethyl sulfoxide (DMSO; 0.5
mL) was heated to reflux for 21 h under argon. The volatiles
were removed under vacuum. The residue was dissolved in
toluene (50 mL) and washed with 5% aqueous acetic acid (25
mL) and brine (25 mL). The organic solution was dried over
NaOH pellets, filtered, and evaporated to dryness. The residue
was purified by column chromatography (silica gel: petroleum
ether/dichloromethane ) 1/2). Recrystallization from dichlo-
romethane/hexanes gave 7f (15 mg, 55%) as a purple solid:
1H NMR δ 9.03 (d, 2H, J ) 5 Hz), 8.98-8.93 (m, 6H), 8.25-
8.21 (m, 6H), 7.84-7.73 (m, 9H), 3.76 (d, 4H, J ) 3.5 Hz),
3.68 (d, 4H, J ) 3.0 Hz); 19F NMR δ 24.20-24.10 (m, 2F),
22.27 (dd, 2F, J ) 9 and 23 Hz), 21.84-21.66 (m, 2F), 11.586
(dd, 2F, J ) 6 and 20 Hz), 10.45 (t, 1F, J ) 22 Hz), 10.12 (dd,
2F, J ) 9 and 23 Hz), 0.82-0.64 (m, 2F); UV λmax 418 (s),
548 (w), 588 (w) nm. Anal. Calcd for C60H31F13N6Zn: C, 62.76;
H, 2.72; N, 7.23. Found: C, 63; H, 2.73; N, 7.14.
1
15/1/0.1) provided 7a (97 mg, 89%) as a purple solid: H NMR
δ 9.05-8.90 (m, 8H), 8.33-8.10 (m, 6H), 7.90-7.64 (m, 9H),
3.89 (m, 4H), 3.78 (m, 5H); UV λmax 418 (s), 548 (w), 588 (w)
nm. MS. Calcd for C48H33N6F4Zn [(M + H)+]: 834. Found:
834.
ZnTPPF4-bip (7c). 4,4′-Bipiperidyl was prepared from its
commercially available dihydrochloride (362 mg, 1.5 mmol)
by treatment with 1 M NaOH (3 mL), to produce a white solid
[229 mg, 91%, sublimation > 150 °C, mp 172 °C (lit.24 172-
173 °C)]: 1H NMR δ 3.02 (d, 4H, J ) 11.7 Hz, H2a), 2.50
(dd, 4H, J ) 11.4 Hz, H2b), 1.80-1.46 (m, 6H) 1.28-0.95
(m, 6H). A solution of zinc porphyrin 2 (190 mg, 0.248 mmol)
and 4,4′-bipiperidyl (125 mg, 0.734 mmol) in freshly distilled
dry 1-methyl-2-pyrrolidinone (3 mL) was heated to 110-120
°C for 18 h under argon. The reaction mixture was then cooled
to room temperature, diluted with dichloromethane (50 mL),
ZnTPPCl6F4-dmdap (7g). 4 (130 mg, 0.133 mmol) was
dissolved in excess N,N′-dimethyl-1,3-diaminopropane (1.5 mL)

10544 J. Phys. Chem. A, Vol. 103, No. 49, 1999
Portela et al.
and stirred for 2 h at 135 °C. After cooling to room temperature,
the mixture was diluted with dichloromethane (50 mL), washed
with a saturated ammonium chloride solution (2 × 20 mL),
water (2 × 20 mL), and brine (2 × 20 mL), dried over
magnesium sulfate, filtered, and evaporated in vacuo to dryness.
The crude product was recrystallized from chloroform/hexane
to yield 7g (137 mg, 97%) as a purple solid: 1H NMR δ 9.01
(d, 2H, J ) 4.5 Hz, pyrrole), 8.75-8.70 (m, 6H, pyrrole), 7.78-
7.57 (m, 9H), 3.68 (t, 2H, J ) 6.9 Hz), 3.61 (m, 2H), 3.36 (bs,
6H), 3.29 (bs, 1H), 2.38-2.30 (m, 2H); 19F NMR (DMSO-d6)
δ 21.37 (bs, 2F), 11.02 (bs, 2F); UV λmax 419 (s), 550 (w), 589
(w) nm. Anal. Calcd for C49H30Cl6F4N6Zn: C, 55.68; H, 2.86;
N, 7.95. Found: C, 54.98; H, 2.81; N, 7.86.
bicarbonate (3 × 10 mL), and brine (3 × 10 mL). The water
layers were combined and washed with dichloromethane (2 ×
40 mL). The combined organic layer was dried over sodium
sulfate, filtered, and evaporated in vacuo, and the residual
1-methyl-2-pyrrolidinone was removed by vacuum distillation
at 60 °C. Purification by column chromatography (silica gel:
dichloromethane/methanol/ammonium hydroxide ) 15/1/0.1)
provided 8b (60 mg, 34%) as a dark purple solid: 1H NMR δ
9.06-8.92 (m, 16H), 8.28-8.20 (m, 12H), 7.84-7.70 (m, 18H),
2.26 (s, 12H); UV λmax 419 (s), 547 (w), 585 (w) nm. MS. Calcd
for C96H61N10F8Zn2 [(M + H)+]: 1633. Found: 1633.
Zn,Zn(TPPF4)2-bip (8c). A solution of zinc porphyrin 7c
(65.6 mg, 0.071 mmol), 2 (100 mg, 0.14 mmol), and tributy-
lamine (0.034 mL, 0.14 mmol) in freshly distilled dry 1-methyl-
2-pyrrolidinone (1 mL) was heated to 110-120 °C for 24 h
under argon. After cooling to room temperature, the mixture
was diluted with dichloromethane (50 mL) and washed with
water (3 × 20 mL), a saturated sodium bicarbonate solution (3
× 20 mL), and brine (3 × 20 mL). The water layers were
combined and rewashed with dichloromethane (2 × 40 mL).
The combined organic layer was dried over sodium sulfate. The
solvent was evaporated in vacuo, and the residual 1-methyl-2-
pyrrolidinone was removed by vacuum distillation at 60 °C.
Purification by column chromatography (silica gel: dichlo-
romethane/methanol ) 50/1) gave 8c (78 mg, 66%) as a purple
solid: 1H NMR δ 9.10-8.90 (m, 8H), 8.25-8.21 (m, 6H),
7.78-7.73 (m, 9H), 3.88-3.83 (m, 2H), 3.51-3.46 (m, 2H),
2.18-2.12 (m, 2H), 2.10-2.02 (m, 2H), 1.81-1.71 (m, 2H),
1.64-1.53 (m, 2H), 1.29-1.23 (m, 2H), 1.03-0.90 (m, 2H),
0.90-0.83 (m, 2H); UV λmax 419 (s), 547 (w), 585 (w) nm.
MS. Calcd for C98H65N10F8Zn2 [(M + H)+]: 1661. Found:
1663, 1665 (most abundant).
ZnTPPF4-bip-CHO (7h). Reaction of ZnTPPF5 2 with 4,4′-
bipiperidyl does not proceed at room temperature in DMF and
is extremely slow at 60 °C. Therefore, a solution of 2 (100 mg,
0.13 mmol) and 4,4′-bipiperidyl (225 mg, 1.1 mmol) in freshly
distilled dry DMF (8 mL) was heated to 140 °C for 24 h under
argon. The reaction mixture was cooled to room temperature,
and DMF was removed by vacuum distillation at 40 °C. The
residue was redissolved in dichloromethane (150 mL) and
washed with water (3 × 30 mL), a saturated sodium bicarbonate
solution (3 × 30 mL), and brine (3 × 30 mL). The water layers
were combined and rewashed with dichloromethane (2 × 40
mL). The combined organic layer was dried over sodium sulfate,
filtered, and evaporated in vacuo to dryness. The crude mixture
was purified by column chromatography (silica gel: dichlo-
romethane/methanol ) 50/1 and then 25/1) to provide 7h (0.112
g, 91%) as a dark purple solid: 1H NMR δ 9.01-8.90 (m, 8H),
8.28-8.24 (m, 6H), 7.85-7.72 (m, 9H), 5.29 (s, 1H, CHO),
3.70 (d, 2H, J ) 11.4 Hz, H2a), 3.30 (dd, 2H, J ) 11.8 Hz,
H2b), 3.02 (d, 1H, J ) 12.6 Hz), 2.84 (d, 1H, J ) 12.9 Hz),
2.49-2.39 (m, 1H), 1.75-1.63 (m, 3H), 1.58-1.40 (m, 3H),
1.39-1.10 (m, 3H), 0.81-0.62 (m, 1H), 0.56-0.35 (m, 1H);
13C NMR δ 159.43 (CHO), 150.55, 150.06, 149.76, 142.95,
134.53, 132.89, 132.19, 131.81, 131.23, 129.94, 127.39, 126.45,
126.39, 52.00 (2 CH2), 46.04 (CH2), 40.86 (CH), 40.31 (CH),
39.41 (CH2), 29.87 (CH2), 29.71 (CH2), 29.25 (CH2), 27.91
(CH2); 19F NMR δ 22.50 (dd, 2F, J ) 9 and 23 Hz), 10.56 (dd,
2F, J ) 9 and 23 Hz); UV λmax 419 (s), 547 (w), 586 (w) nm.
MS. Calcd for C55H43F4N6OZn [(M + H)+]: 943. Found: 943.
Zn,Zn(TPPF4)2-pip (8a). A mixture of 2 (19.7 mg, 25 µmol),
7a (64 mg, 77 µmol), and tributylamine in dry DMSO (0.25
mL) was heated to 120 °C for 24 h. After cooling to room
temperature, the solvent was removed by vacuum distillation
at 60-80 °C. The residue was dissolved in dichloromethane
(50 mL), washed with a saturated sodium bicarbonate solution
(2 × 20 mL) and brine (3 × 20 mL), and dried over sodium
sulfate. The solvent was evaporated in vacuo. The residue was
purified by column chromatography (silica gel: dichloromethane/
petroleum ether ) 3/1). The product was recrystallized from
dichloromethane/hexanes to yield 8a (13 mg, 33%) as purple
crystals: 1H NMR δ 9.07 (d, 4H, J ) 4.5 Hz), 9.02 (d, 4H, J
) 4.5 Hz), 8.97 (m, 8H), 8.27-8.22 (m, 12H), 7.85-7.70 (m,
18H), 3.94 (s, 8H); UV λmax 419 (s), 547 (w), 587 (w) nm.
MS. Calcd for C92H55N10F8Zn2 [(M + H)+]: 1579. Found: 1582
(most abundant).
Zn,Zn(TPPF4)2-stda (8e). 5R-Androstane-3â,17â-diamine
was prepared by oxidation of epiandrosterone followed by
reductive amination.26 A solution of this amine (5.2 mg, 0.018
mmol), 2 (30 mg, 0.039 mmol), and tributylamine (0.013 mL,
0.053 mmol) in freshly distilled dry 1-methyl-2-pyrrolidinone
(1 mL) was heated to 140-150 °C for 72 h under argon. The
reaction mixture was worked up in the usual manner. Purifica-
tion by column chromatography (silica gel: dichloromethane/
petroleum ether ) 1/1) gave 8e (7.5 mg, 23%) as a purple
solid: 1H NMR δ 9.30-8.88 (m, 16H), 8.25-8.19 (m, 12H),
7.82-7.72 (m, 18H), 4.65-4.54 (m, 1H), 4.00-3.94 (m, 1H),
2.50-0.50 (m, 28H); UV λmax 418 (s), 548 (w), 586 (w) nm.
MS. Calcd for C107H79N10F8Zn2 [(M + H)+]: 1783. Found:
1783.
Zn,Zn(TPPF4)2-pip-pfbp-pip (8f). A mixture of 7f (10 mg,
0.0087 mmol), 7a (15 mg, 0.018 mmol), 1-ethyl-2,2,6,6-
tetramethylpiperidine (0.1 mL), and dry DMSO (0.5 mL) was
heated to 120 °C for 44 h under argon. The mixture was cooled
to room temperature, diluted with toluene (50 mL), and washed
with 5% aqueous acetic acid (2 × 25 mL), saturated sodium
bicarbonate (2 × 25 mL), and brine (2 × 20 mL). The organic
solution was dried over sodium sulfate, filtered, and evaporated
to dryness. The residue was purified by column chromatography
(silica gel: petroleum ether/dichloromethane ) 1/1). Recrys-
tallization from dichloromethane/hexanes gave 8f (10 mg, 59%)
as a purple solid: 1H NMR δ 9.03 (d, 4H, J ) 5.0 Hz), 8.98-
8.94 (m, 12H), 8.26-8.20 (m, 12H), 7.82-7.73 (m, 18H), 3.82
(d, 8H, J ) 3.5 Hz), 3.68 (d, 8H, J ) 3.0 Hz); 19F NMR δ
22.20 (dd, 4F, J ) 9 and 23 Hz), 21.64 (dd, 4F, J ) 5, 19 Hz),
11.21 (bd, 4F, J ) 15 Hz), 10.11 (dd, 4F, J ) 9, 23 Hz); UV
Zn,Zn(TPPF4)2-bcoda (8b). A solution of zinc porphyrin 2
(205 mg, 0.267 mmol), bicyclo[2.2.2]octane-1,4-diamine25 (15
mg, 0.107 mmol), and tributylamine (0.64 mL, 0.267 mmol) in
freshly distilled dry 1-methyl-2-pyrrolidinone (3 mL) was heated
to 150-160 °C for 48 h under argon. The reaction mixture was
cooled to room temperature, diluted with dichloromethane (50
mL), and washed with water (3 × 10 mL), a saturated sodium

Electron Transfer in Metalloporphyrin Dimers
max 417 (s), 549 (w), 586 (w) nm. Anal. Calcd for C108H62F16N12-
J. Phys. Chem. A, Vol. 103, No. 49, 1999 10545
λ
decanted, the solvent was evaporated, and the residue was
Zn2: C, 66.10; H, 3.18; N, 8.56. Found: C, 65.91; H, 3.36; N,
7.96.
recrystallized from methylene chloride/hexanes to yield 9f (40
mg, 88%) as purple crystals. H NMR δ 9.00 (d, 4H, J ) 5.1
1
Hz), 8.98-8.91 (m, 12H), 8.27-8.18 (m, 12H), 7.84-7.74 (m,
18H), 3.80 (d, 8H, J ) 3.5 Hz), 3.69 (d, 8H, J ) 3.1 Hz), -2.73
(s, 4H); UV λmax 418 (s), 513 (w), 547 (w), 587 (w), 643 (w)
nm. Anal. Calcd for C108H66F16N12: C, 70.66; H, 3.62; N, 9.16.
Found: C, 70.79; H, 3.55; N, 8.97.
Zn,Zn(TPPCl6F4)2-dmdap (8g). Amine 7g (62.0 mg, 0.059
mmol), 4 (28.6 mg, 0.029 mmol), and tributylamine (6.9 µL,
0.029 mol) were dissolved in dry DMSO (300 µL) under an
argon atmosphere and stirred at 130 °C for 10 h. After cooling
to room temperature, the solution was diluted with dichlo-
romethane (100 mL). The organic phase was washed with a
saturated ammonium chloride solution (3 × 10 mL) and brine
(3 × 10 mL). After drying over magnesium sulfate, the solvent
was evaporated in vacuo. The crude mixture was purified by
column chromatography (silica gel: petroleum ether/dichlo-
romethane ) 3:1) to give 8g (32 mg, 54% based on 4) as a
purple solid: 1H NMR δ 8.96 (d, 4H, J ) 4.5 Hz, pyrrole),
8.74 (d, 4H, J ) 4.5 Hz, pyrrole), 8.73(d, 4H, J ) 4.5 Hz,
pyrrole), 8.70 (d, 4H, J ) 4.5 Hz, pyrrole), 7.76 (d, 4H, J ) 8
Hz, m), 7.71-7.66 (m, 10H, J ) 8 Hz, phenyl), 7.57 (t, 4H, J
) 8 Hz, p), 3.71 (t, 4H, J ) 6.8 Hz, N-CH2), 3.34 (s, 6H,
CH3), 2.31 (qn, 2H, J ) 6.8 Hz, CH2); 19F NMR δ 22.27 (d,
4F, J ) 16 Hz), 10.50 (d, 4F, J ) 16 Hz); UV λmax (10-4ꢀ) 419
(26), 549 (1.6), 585 (sh) nm. MS (FAB+, CH2Cl2/NBA). Calcd
for C93H47Cl12F8N10Zn2 [(M + H)+]: 2003. Found: 2012 (most
abundant).
H2,H2(TPPCl6F4)-dmdap (9g). Dimer 8g (21 mg, 0.0104
mmol) was dissolved in dichloromethane (30 mL) and mixed
with 25% aqueous hydrochloric acid (5 mL). After 12 h of
reflux, a brownish green solution was obtained. Two phases
were separated. The dichloromethane layer was washed with
water until the pH was neutral, whereupon the organic layer
changed to dark red. The solution was then dried over sodium
sulfate, concentrated by evaporation in vacuo, and purified by
column chromatography (silica gel: petroleum ether/dichlo-
romethane ) 1/1) to obtain 9g (16 mg, 80%) as a purple solid:
1H NMR δ 8.90 (d, 4H, J ) 4.5 Hz, pyrrole), 8.69 (d, 4H, J )
4.5 Hz, pyrrole), 8.68 (d, 4H, J ) 4.5 Hz, pyrrole), 8.65 (d,
4H, J ) 4.5 Hz, pyrrole), 7.79 (d, 4H, J ) 8 Hz, m), 7.69 (t,
2H, J ) 8 Hz, p), 7.67 (m, 8H, J ) 8 Hz, m), 7.53 (t, 4H, J )
8 Hz, p), 3.71 (t, 4H, J ) 7.5 Hz, N-CH2), 3.35 (s, 6H, CH3),
2.33 (qn, 2H, J ) 7.5 Hz, CH2), -2.60 (s, 4H, NH); 19F NMR
δ 22.47 (dd, 4F, J ) 23 and 8 Hz), 10.63 (dd, 4F, J ) 23 and
8 Hz); UV λmax (10-4ꢀ) 418 (61), 512 (4.2), 543 (0.6), 588 (1.3),
645 (0.1) nm. Anal. Calcd for C93H50Cl12F8N10: C, 59.26; H,
2.67; N, 7.73. Found: C, 59.08; H, 2.69; N, 7.59.
H2,H2(TPPF4)2-pip (9a). A solution of 8a (15 mg, 9.5 mmol)
in dichloromethane (15 mL) was vigorously stirred with 25%
aqueous hydrochloric acid (10 mL) for 20 h. The organic layer
was separated, diluted with dichloromethane (50 mL), washed
with water (3 × 15 mL), 10% NaOH (2 × 15 mL), and brine
(3 × 15 mL), and dried over NaOH pellets. The solution was
decanted, the solvent was evaporated, and the residue was
recrystallized from dichloromethane/hexanes to yield 9a (13 mg,
87%) as purple crystals: 1H NMR δ 8.96 (d, J ) 4.5 Hz, 4H),
8.92 (d, J ) 4.5 Hz, 4H), 8.93-8.31 (m, 8H), 8.27-8.22 (m,
12H), 7.84-7.74 (m, 18H), 3.93 (s, 8H), -2.73 (s, 4H); UV
λmax 418 (s), 513 (w), 548 (w), 588 (w), 643 (w) nm. Anal.
Calcd for C92H58F8N10: C, 75.92; H, 4.02; N, 9.62. Found: C,
76.19; H, 3.99; N, 9.39.
Zn,H2(TPPF4)2bip (10c). A solution of zinc porphyrin 7c
(65.6 mg, 0.071 mmol), 1 (100 mg, 0.14 mmol), and tributyl-
amine (0.034 mL, 0.14 mmol) in freshly distilled dry 1-methyl-
2-pyrrolidinone (1 mL) was heated to 110-120 °C for 24 h
under argon. The reaction mixture was cooled to room tem-
perature, diluted with dichloromethane (50 mL), and washed
with water (3 × 20 mL), a saturated sodium bicarbonate solution
(3 × 20 mL), and brine (3 × 20 mL). The water layers were
combined and rewashed with dichloromethane (2 × 40 mL).
The combined organic layer was dried over sodium sulfate,
filtered, and evaporated in vacuo, and the residual 1-methyl-2-
pyrrolidinone was removed by vacuum distillation at 60 °C.
Purification by column chromatography (silica gel: dichlo-
romethane/methanol ) 50/1) gave 10c (78 mg, 66%) as a purple
solid: 1H NMR δ 9.05-8.49 (m, 16H), 8.26-8.21 (m, 12H),
7.82-7.75 (m, 18H), 3.90-3.85 (m, 2H), 3.52-3.42 (m, 2H),
2.10-2.02 (m, 2H), 1.77-1.72 (m, 2H), 1.62-1.55 (m, 1H),
1.55-1.44 (m, 3H), 1.32-1.21 (m, 3H), 0.93-0.75 (m, 3H),
-2.69 (s, 2H); UV λmax 419 (s), 510 (w), 548 (w), 585 (w),
622(w) nm. HRMS. Calcd for C98H67N10F8Zn [(M + H)+]:
1599.47. Found: 1599.6.
H2,H2(TPPF4)2-bcoda (9b). A mixture of 8b (20 mg, 12.2
mmol), dichloromethane (20 mL), methanol (2 mL), and
aqueous hydrochloric acid (1M, 10 mL) was stirred for 10 h at
room temperature. The reaction mixture was worked up ac-
cording to the procedure for 9a. After evaporation of the solvent,
the residue was taken up in CH2Cl2 and precipitated with
pentane to yield 9b (15 mg, 83%) as a dark purple solid: 1H
NMR δ 8.98-8.92 (m, 16H), 8.18-8.28 (m, 12H), 7.82-7.72
(m, 18H), 2.30 (s, 12H), -2.75 (s, 4H); UV λmax 418 (s), 514
(w), 549 (w), 591 (w), 644 (w) nm.
H2,H2(TPPF4)2-stda (9e). A mixture of 8e (4 mg, 0.002
mmol) in dichloromethane (5 mL) and methanol (0.5 mL) was
treated with aqueous hydrochloric acid (1 M, 1 mL) for 12 h,
followed by the usual workup. After evaporation of the solvent,
the residue was taken up in CH2Cl2 and precipitated with
pentane to yield 9e (3 mg, 91%) as a dark purple solid: 1H
NMR δ 9.27-8.82 (m, 16H), 8.28-8.15 (m, 12H), 7.80-7.70
(m, 18H), 4.50-4.40 (m, 1H), 3.90-3.70 (m, 1H), 2.50-0.50
(m, 28H), -2.70 (s, 4H); UV λmax 417 (s), 513 (w), 548 (w),
589 (w), 643 (w) nm.
Zn,H2(TPPF4)2stde (10d). Solid 7d (3 mg, 2.8 µmol) was
added to a mixture of sodium hydride (0.53 mg, 0.013 mmol,
60% dispersion in paraffin) and 1-methyl-2-pyrrolidinone (1 mL)
at 5 °C under argon. After 1 h 1 (10 mg, 0.014 mmol) was
added, and the mixture was heated to 50 °C for 48 h. After
cooling to room temperature, the mixture was poured onto ice
water (50 mL) and washed with dichloromethane (3 × 15 mL).
The combined organic layer was dried over sodium sulfate and
evaporated to dryness. Purification by column chromatography
(silica gel: petroleum ether/dichloromethane ) 1/1) gave 10d
(1.9 mg, 38%) as a purple solid: 1H NMR δ 9.10-8.80 (m,
16H), 8.20-8.10 (m, 12H), 7.80-7.70 (m, 18H), 5.00-4.90
(m, 1H), 4.90-4.80 (m, 1H), 2.60-0.50 (m, 28H), -2.70 (s,
2H); UV λmax 419 (s), 510 (w), 548 (w), 585 (w), 620(w) nm.
H2,H2(TPPF4)2-pip-pfbp-pip (9f). A solution of 8f (50 mg,
0.025 mmol) in dichloromethane (25 mL) was vigorously stirred
with 25% aqueous HCl (20 mL) for 20 h. The organic layer
was separated, diluted with dichloromethane (50 mL), washed
with water (3 × 20 mL), 10% NaOH (2 × 20 mL), and brine
(3 × 20 mL), and dried over NaOH pellets. The solution was

10546 J. Phys. Chem. A, Vol. 103, No. 49, 1999
Portela et al.
Zn,Fe(TPPF4)2-pip‚OAc (12a). A mixture of dimer 9a (8.5
mg, 5.8 µmol) and ferrous bromide (1.25 mg, 5.8 µmol) in
freshly distilled THF (15 mL) was heated to reflux for 1.5 h
under argon. At this time additional ferrous bromide (1.0 mg,
4.6 µmol) was added and heating was continued for 24 h. After
cooling to room temperature, the reaction mixture was exposed
to air for 30 min to allow the spontaneous oxidation of iron(II)
to iron(III), and the solvent was evaporated in vacuo. The residue
was dissolved in dichloromethane (30 mL), washed with a 10%
NaOH solution (2 × 10 mL), and dried over NaOH pellets.
The solution was decanted, and the solvent was evaporated in
vacuo. The crude mixture was purified by column chromatog-
raphy (silica gel: dichloromethane/methanol ) 100/1) to give
monometalated dimer 11a (8.3 mg, 60%) as a purple-brown
solid. Without further characterization monometalated dimer 11a
was dissolved in a dichloromethane/methanol ) 2/1 solution
(10 mL). Zinc acetate dihydrate (100 mg, 0.045 mmol) was
added, and the mixture was heated to reflux for 2 h. After
cooling to room temperature, the solvent was evaporated in
vacuo. The residue was dissolved in dichloromethane (25 mL),
washed with water (2 × 15 mL) and 5% aqueous acetic acid (3
× 15 mL), and dried over sodium sulfate. After evaporation of
the solvent, the residue was purified by column chromatography
(silica gel: 1% methanol in dichloromethane) to yield the single
species 12a (5.5 mg, 58%): UV λmax 418 (s), 509 (w), 547 (w)
nm. MS. Calcd for C94H60N10F8O2FeZn [(M + H)+]: 1632.
Found: 1632.
purified by column chromatography (silica gel: dichloromethane/
methanol ) 50/1) to yield 12d (1.9 mg, 70%): UV λmax 418
(s), 511 (w), 548 (w) nm. MS. Calcd for C109H81N8F8O4FeZn
[(M + H)+]: 1837. Found: 1777 [(M-OAc)+].
Zn,Fe(TPPF4)2-stda‚OAc (12e). A mixture of dimer 9e (2.2
mg, 1.3 µmol) and ferrous bromide (1.4 mg, 6.6 µmol) in freshly
distilled dry THF (1 mL) was heated to reflux for 10 h under
argon. After workup and zinc acetate treatment, as for 12b, the
crude mixture was purified by column chromatography (silica
gel: dichloromethane/methanol ) 50/1) to yield 12e (1.5 mg,
65%, mixture of regioisomers): UV λmax 418 (s), 513 (w), 547
(w) nm. MS. Calcd for C109H82N10F8O2FeZn [(M + H)+]: 1834.
Found: 1775 (M + H+ - OAc).
Zn,Fe(TPPF4)2-pip-pfbp-pip‚OAc (12f). A mixture of dimer
9f (10 mg, 0.005 mmol) and ferrous bromide (1.25 mg, 5.8
µmol) in freshly distilled THF (15 mL) was heated to reflux
under argon for 10 h. After the usual workup, the crude mixture
was purified by column chromatography (silica gel: dichlo-
romethane/methanol ) 2/1) to give monometalated dimer 11f
(5.7 mg, 59%) as a purple-brown solid. Without further
characterization, monometalated dimer 11c was dissolved in 10
mL of 1/1 dichloromethane/methanol. Zinc acetate dihydrate
(100 mg, 0.045 mmol) was added, and the mixture was stirred
at 25 °C for 1.5 h. The reaction was briefly heated to reflux
(5-10 min), and the solvent was evaporated in vacuo. The
residue was dissolved in dichloromethane (25 mL), washed with
water (2 × 15 mL) and 5% aqueous acetic acid (3 × 15 mL),
and dried over sodium sulfate. After evaporation of solvent,
the residue was purified by column chromatography (silica
gel: 1% methanol in dichloromethane) to yield 12f (5.5 mg,
58%): UV λmax 418 (s), 510 (w), 547 (w) nm. MS. Calcd for
Zn,Fe(TPPF4)2-bcoda‚OAc (12b). A mixture of dimer 9b
(4.5 mg, 3.0 µmol) and ferrous bromide (0.65 mg, 3.0 µmol) in
freshly distilled THF (1 mL) was heated to reflux for 24 h under
argon. After the usual workup, the crude mixture was purified
by column chromatography (silica gel: dichloromethane/
methanol ) 50/1) to give monometalated dimer 11b (3 mg,
64%) as a purple-brown solid. Without further characterization,
monometalated dimer 11b and zinc acetate dihydrate (50 mg,
0.02 mmol) were heated to reflux in dichloromethane/methanol
) 1/1 (5 mL) for 2 h. After cooling to room temperature, the
solvent was evaporated in vacuo. The residue was dissolved in
dichloromethane (25 mL), washed with water (2 × 15 mL) and
5% aqueous acetic acid (3 × 15 mL), and dried over sodium
sulfate. After evaporation of the solvent, the residue was purified
by column chromatography (silica gel: 1% methanol in dichlo-
romethane) to yield 12b (2 mg, 65%): UV λmax 418 (s), 512
(w), 547 (w) nm. MS. Calcd for C98H64N10F8O2FeZn [(M +
H)+]: 1684. Found: 1684.
C
110H66N12F16O2FeZn [(M + H)+]: 2010. Found: 2012 (most
abundant).
Zn,Fe(TPPCl6F4)2-dmdap‚OAc (12g). A mixture of dimer
9g (14.5 mg, 7.7 µmol) and ferrous bromide (1.8 mg, 8.4 µmol)
in freshly distilled DMF (2.5 mL) was heated to reflux under
argon for 20 min. After this time, additional ferrous bromide
(0.5 mg, 2.3 µmol) was added and reflux was continued for 10
min. The reaction mixture was diluted with dichloromethane
(50 mL) and washed with water (3 × 20 mL). The solvent was
evaporated in vacuo. The crude mixture was purified by column
chromatography (alumina: dichloromethane/methanol ) 1/1)
to give monometalated dimer 11g (8.3 mg, 63%) as a purple-
brown solid. Without further characterization, this was dissolved
in 1/1 dichloromethane/methanol (10 mL). Zinc acetate dihy-
drate (172.2 mg, 0.077 mmol) was added, and the mixture was
heated to reflux for 2 h. After cooling to room temperature, the
solvent was evaporated in vacuo and the residue was dissolved
in dichloromethane (50 mL), washed with 5% aqueous acetic
acid (3 × 20 mL) and water (3 × 20 mL), and dried over sodium
sulfate. The crude mixture was purified by gradient column
chromatography (alumina: dichloromethane, gradually to dichlo-
romethane/methanol ) 50/1) to give dimer 12g (5.5 mg, 63%)
as a dark purple solid: UV λmax 417 (s), 510 (w), 549 (w) nm.
MS. Calcd for C95H52Cl12F8N10O2FeZn [(M + H)+]: 2055.
Found: 2061 (most abundant).
Zn,Fe(TPPF4)2-bip‚OAc (12c). A mixture of dimer 10c (66
mg, 0.041 mmol) and ferrous bromide (17.4 mg, 0.078 mmol)
in freshly distilled dry THF (5 mL) was heated to reflux for 2
h under argon. After cooling to room temperature, the reaction
mixture was aerated for 30 min. The reaction mixture was then
diluted with dichloromethane (60 mL) and washed with 5%
aqueous acetic acid (3 × 20 mL) and water (3 × 20 mL). The
water layers were combined and rewashed with dichloromethane
(2 × 20 mL). The combined organic layer was dried over
sodium sulfate, filtered, and evaporated in vacuo to dryness.
The crude mixture was purified by column chromatography
(silica gel: dichloromethane/methanol ) 50/1) to yield 12b
(0.064 g, 88%): UV λmax 418 (s), 513 (w), 548 (w) nm. MS.
Calcd for C100H69N10F8O2FeZn [(M + H)+]: 1713. Found:
1713.
Results and Discussion
Considerations of Synthesis. A large part of the effort in
the studies of covalently linked donor-acceptor systems lies
in the skillful synthesis of tailor-made supramolecular struc-
tures.8,10,13 Dimeric and oligomeric porphyrin arrays have been
prepared by several different building-block approach-
es.10,16,18-20,27-30
Zn,Fe(TPPF4)2-stde‚OAc (12d). A mixture of dimer 10d (2.5
mg, 1.5 µmol) and ferrous bromide (1.56 mg, 7.3 µmol) in
freshly distilled dry THF (2 mL) was heated to reflux for 12 h
under argon. After the usual workup, the crude mixture was

Electron Transfer in Metalloporphyrin Dimers
J. Phys. Chem. A, Vol. 103, No. 49, 1999 10547
SCHEME 1a
a (a) BF3‚OEt2, 2,2-dimethoxypropane, CH2Cl2, rt, 2 h. (b) DDQ, rt, 5 h, 20-30%. (c) FeBr2, DMF or THF, ∆. (d) Aqueous AcOH. (e)
Zn(OAc)2‚2H2O, CH2Cl2/MeOH, ∆, 98%.
Our synthetic strategy represents an application of a regi-
oselective nucleophilic aromatic substitution previously de-
scribed in porphyrin chemistry.20 The regioselectivity of the
coupling depends on a porphyrin bearing a single meso-
pentafluorophenyl group, where only the fluorine at the para
position can be successfully substituted with a nitrogen nucleo-
phile. This synthetic methodology allows the systematic modi-
fication of many variables in the diporphyrin system, including
type and length of spacer, metal center, and redox-potential
difference between the donor and acceptor metalloporphyrin.
We prepared a set of building blocks suitable for the sequential
synthesis of all of the porphyrin dimers in Figure 1.
The reaction scheme calls for a suitable porphyrin with a
single meso-pentafluorophenyl group (Scheme 1). Stable,
crystalline H2TPPF5 (1) fulfills the basic demands. Its synthesis
was accomplished by a modification of the procedure of
Lindsey.31,32 A 7/1 mixture of benzaldehyde and pentafluo-
robenzaldehyde was condensed with 1 equiv of pyrrole under
Lewis acid catalysis (BF3‚Et2O) and 2,2-dimethoxypropane as
a water scavenger, to provide a presumed porphyrinogen
intermediate that was not isolated. Oxidative aromatization with
DDQ then provided the desired porphyrin in respectable yield.
Pentafluoroporphyrin 1 is readily metalated with excess zinc
acetate to form zinc(II) porphyrin 2. Because this is stable under
basic conditions, it proved to be suitable for nucleophilic
substitution with amines. To demonstrate the generality of this
method, 2,6-dichlorobenzaldehyde and pentafluorobenzaldehyde
were also coupled to form tris(2,6-dichlorophenyl)pentafluo-
roporphyrin 3. Metalation of 3 provided zinc porphyrin 4 in
excellent yield. The iron(III) porphyrins FeTPPF5‚OAc (5) and
FeTPPCl6F5‚OAc (6) could also be prepared by a standard iron
insertion method.33,34
Heating 2 with excess piperazine in either DMF or 1-methyl-
2-pyrrolidinone leads to clean substitution of the para fluorine
and conversion to amino derivative 7a (Scheme 2). This amine
was next heated in DMSO with an excess of 2 and tributylamine
as a proton scavenger to afford Zn,Zn dimer 8a in good yield.
This two-step procedure, using an excess of one reagent and
then the other, was found to be preferable to a single reaction
under 2/1 stoichiometry. Hydrolysis of the Zn,Zn dimer 8a with
HCl in methanol/dichloromethane, followed by neutralization,
provided free base dimer 9a. Iron insertion into 9a was achieved
with 1.1 equiv of FeBr2 in refluxing THF, followed by a standard
ligand exchange and air oxidation, to provide monometalated
iron(III) acetate dimer 11a. Dimer 11a underwent zinc insertion
to form the donor-acceptor dimer 12a in 30% yield from 8a.
The Zn,Zn dimer 8b was obtained in only 34% yield by a
one-step double coupling of 2 with bicyclo[2.2.2]octane-1,4-
diamine25a and tributylamine in 1-methyl-2-pyrrolidinone. Acid
hydrolysis provided 9b, which was converted to 12b by
sequential treatment with ferrous bromide and zinc acetate.
A different result was obtained when this coupling was
attempted toward 8c. Heating 2 with 4,4′-bipiperidyl in DMF
proceeds with clean substitution of the aromatic fluoride.
However, subsequent attack of the remaining amine functionality
on DMF results in transamidation and formation of formamide
7h as the sole product. This amide was found to be stable to
various reaction conditions and unreactive in additional coupling
reactions. When DMF was replaced with 1-methyl-2-pyrroli-
dinone, which is less prone to transamidation, the reaction of 2
with excess 4,4′-bipiperidyl gave exclusively 7c. This material
showed interesting NMR behavior. The aliphatic CH signals
of the 4,4′-bipiperidyl fragment are broad and show an extreme
upfield shift, to δ -2.8 ppm in CDCl3. Such broadening was

10548 J. Phys. Chem. A, Vol. 103, No. 49, 1999
Portela et al.
SCHEME 2a
a (a) Excess piperazine, 1-methyl-2-pyrrolidinone, 120 °C, 24 h, 96-98%. (b) ZnTPPF5, tributylamine, 1-methyl-2-pyrrolidinone, 130 °C, 24 h,
87%. (c) 1 M HCl in MeOH/CH2Cl2. (d) FeBr2, THF, ∆. (e) Zn(OAc)2‚2H2O, CH2Cl2/MeOH, ∆, 92%.
not observed in formamide 7h. At 50 °C some of the multiplets
of 7c sharpen, suggesting a dynamic complexation and decom-
plexation on the NMR time scale. We propose that the free
amino group is cooordinated to the zinc of another molecule,
producing a “head to tail” dimer, where the aliphatic protons
are shielded by the electron cloud of the second porphyrin. The
Zn,Zn dimer 8c was then obtained by coupling amine 7c with
excess 2. Alternatively, 7c could be coupled with an excess of
1 to yield monozinc dimer 10c. Iron insertion into the free-
base porphyrin unit of 10c was effected with ferrous bromide,
followed by acetic acid wash and spontaneous oxidation of Fe-
(II), leading to the formation of Zn(II),Fe(III) acetate bipiperidyl
dimer 12c in good overall yield.
excess perfluorobiphenyl gave 7f, which after a second coupling
with a slight excess of 7a provided Zn,Zn dimer 8f. This dimer
underwent acid-catalyzed zinc hydrolysis, iron insertion, and a
series of ligand-exchange steps to produce 12f.
Zinc porphyrin 4 was reacted with excess N,N′-dimethyl-1,3-
diaminopropane to form amine 7g. When 7g was coupled with
half an equivalent of 4, the Zn,Zn dimer 8g was obtained. This
dimer could be converted in good overall yield to the Zn(II),-
Fe(III) dimer 12g by acid removal of the zinc to produce 9g,
followed by iron insertion, spontaneous oxidation, zinc insertion,
and ligand exchange.
Kinetics of PET. Fluorescence lifetimes were measured for
the series of metalloporphyrin heterodimers 12a-g as well as
for some of their reference homodimers 8a-g. Figure 2 shows
representative fluorescence decay curves. Excited-state lifetimes
determined from these measurements are presented in Table 1.
The fluorescence decays for nearly all of the porphyrin dimers
are monoexponential or else biexponential with a slower
component of very low amplitude. A few of the Zn,Fe dimers
are exceptions. The Zn,Fe dimer 12a linked by piperazine (Table
1, entry 5) exhibits triexponential kinetics with one dominant
component. Two minor components with slower decays were
present, attributable to trace impurities. The lifetime of the third
component is similar to that of the Zn,Zn dimer, suggesting
the presence of ca. 2% of this compound as well. The Zn,Fe
porphyrin dimers 12c and 12f (Table 1, entries 9 and 15)
routinely exhibit a more pronounced biexponential kinetics. The
faster component may be assigned to the Zn,Fe porphyrin, and
the slower, of low amplitude (7% or 15%), can be attributed to
an impurity or decomposition product. When the two chro-
mophores are connected by the flexible propanediamine linker,
dimer 12g exhibits more complex kinetics that required a
5R-Androstane-3R,17â-diol was deprotonated with NaH and
coupled with 2 to form a mixture of Zn-stde monoporphyrin
1
7d (19%) and Zn,Zn dimer 8d (40%). By comparison of H
NMR spectra, the structure of 7d is consistent with a free 3R-
hydroxy derivative, which agrees with the greater reactivity of
17â-diol toward esterification.35 The alcohol 7d was then added
to excess NaH in 1-methyl-2-pyrrolidinone and heated with 1
to produce Zn,H2 dimer 10d. Reaction of 10d with ferrous
bromide led to Zn(II),Fe(III) dimer 12d. Regardless of the
assignments for 7d, 10d, and 12d, the distance between the two
porphyrins is independent of the regiochemistry.
The Zn,Zn-stda dimer 8e was obtained by heating 5R-
androstane-3â,17â-diamine26 with 2 equiv of 2. The mixture
of regioisomeric Zn,Fe-stda dimers 12e could then be obtained
by acid hydrolysis to 9e followed by two successive metal
insertions. Again the distance between the two porphyrins is
the same for both regioisomers.
The Zn,Fe-piperazyl-octafluorobiphenyl-piperazyl dimer 12f
was made analogously to Scheme 2. Heating amine 7a with

Electron Transfer in Metalloporphyrin Dimers
J. Phys. Chem. A, Vol. 103, No. 49, 1999 10549
the same in both regioisomers. Therefore, our choice of edge-
to-edge distance, rather than the more variable metal-metal
distance, is justified, despite the possibility of internal rotations.
The measurements were generally conducted under conditions
such that the acetate ligand on the iron was replaced in situ by
two N-methylimidazoles. The Fe(III) acetate complex, with a
high-spin iron located above the plane of the porphyrin, was
thereby converted to an octahedral bis(imidazole)Fe(III) cationic
complex with a low-spin iron in the porphyrin plane.36 Table 1
shows that dimer 12a, in the absence of 1-methylimidazole
(entry 4), has a longer fluorescence lifetime than that in the
presence of this ligand (entry 5). This result is consistent with
faster ET to the low-spin planar six-coordinated bis(1-meth-
ylimidazole)iron(III) than to the high-spin five-coordinated iron-
(III) that is above the porphyrin plane.
The series of dimers 12a-g represents a selection of linkers.
The lifetimes in Table 1 generally increase with increasing
length of the linker, from 0.1 and 0.3 ns for 12a and 12b, with
one ring besides the C6F4 units, through 0.36 ns for 12c, with
two rings, to 0.72, 0.85, and 1.4 ns for 12d, 12e, and 12f, with
four rings. The three-fold through-bond pathway for electron
transfer through the bicyclo[2.2.2]octane link of 12b does not
overcome the shorter distance in 12a. The steroidal dimers 12d
and 12e show quite similar lifetimes, indicating that there is no
substantial difference between oxygen and nitrogen linkages of
ether and amine.
Figure 2. Typical 650-nm fluorescence decay curves for porphyrin
dimers in dichloromethane. Dots are data points at 16.26-ps intervals;
continuous lines are fits (which often overlay the dots). The curves are
convoluted with the instrument response and are scaled by small factors
for display on a common axis, in order of maxima from left to right:
Instrument response function. 12c: fit to 0.927 exp(-t/0.361 ns) +
0.073 exp(-t/1.32 ns). 9c: fit to 1.0 exp(-t/1.60 ns).
TABLE 1: Fluorescence Lifetimes of Monomeric and
Dimeric Porphyrins in Dichloromethane
%
entry
compounda
structure fitting
τ, ns
Of particular interest is the comparison of the fluorescence
lifetimes observed for the Zn,Fe heterodimers 12a-f (Table 1)
with those of the reference Zn,Zn homodimers 8a-f. The
lifetime of the homodimer provides the rate for all of the intrinsic
deactivation processes in each skeleton under study. Actually
there is little influence of the distant skeleton, because all of
the TPPF4 homodimers 8a-f have nearly the same lifetime,
which is quite close to that of the Zn monomer 2, and dimer 8g
has the same lifetime as monomer 4. The comparison of
heterodimer with homodimer then provides a measure of ET,
which represents the dominant fluorescence quenching mechan-
ism.16a,19 Moreover, the invariance of the observed rates to
variation in dimer concentration from 10 to 100 µM excludes
bimolecular contributions to the observed rates. Therefore, the
deactivation of the photoexcited metalloporphyrin dimers can
be analyzed using eqs 1-3 and converted to electron-transfer
1
2
3
4
5
6
7
8
9
H2TPPF5
ZnTPPF5
Zn,Zn(TPPF4)2-pip
[Zn,Fe(TPPF4)2-pip]‚OAc
[Zn,Fe(TPPF4)2-pip]‚OAcc
Zn,Zn(TPPF4)2-bcoda
[Zn,Fe(TPPF4)2-bcoda]‚OAc
Zn,Zn(TPPF4)2-bip
1
2
8a
100 9.1
99 1.68 ( 0.05
100 1.60 ( 0.05
96b 0.070 ( 0.005
97d 0.107 ( 0.005
100 1.53 ( 0.02
99 0.30 ( 0.02
100 1.60 ( 0.02
93e 0.36 ( 0.02
99 1.57 ( 0.02
98f 0.72 ( 0.02
99 1.59 ( 0.02
99g 0.85 ( 0.02
100 1.59 ( 0.05
85h 1.40 ( 0.05
98f 0.620 ( 0.04
99g 0.620 ( 0.04
57i 0.34
12a
12a
8b
12b
8c
12c
8d
12d
8e
[Zn,Fe(TPPF4)2-bip]‚OAc
10 Zn,Zn(TPPF4)2-stde
11 [Zn,Fe(TPPF4)2-stde]‚OAc
12 Zn,Zn(TPPF4)2-stda
13 [Zn,Fe(TPPF4)2-stda]‚OAc
14 Zn,Zn(TPPF4)2-pip-pfbp-pip
12e
8f
15 [Zn,Fe(TPPF4)2-pip-pfbp-pip]‚OAc 12f
16 ZnTPPCl6F5
4
8g
12g
17 Zn,Zn(TPPCl6F4)2-dmdap
18 [Zn,Fe(TPPCl6F4)2-dmdap]‚OAc
a Abbreviations: pip ) piperazyl; bcoda ) bicyclo[2.2.2]octane-
1,4-diamine; bip ) bipiperidyl; dmdap ) N,N′-dimethyl-1,3-diamino-
propyl; pfbp ) perfluorobiphenyl; stda ) steroidal diamine (5R-
androstanyl-3â,17â-diamine); stde ) steroidal diether (5R-androstanyl
3R,17â-diether). b Also 2% each with τ ) 0.35 ( 0.07 and 1.6 ( 0.1
ns. c Without 1-methylimidazole. d Also 3% with τ ) 2 ns. e Also 7%
with τ ) 1.32 ( 0.02 ns. f Also 2% with τ ) 2 ns. g Also 1% with τ
) 2 ns. h Also 15% with τ ) 0.65 ( 0.05 ns. i Also 38% with τ )
0.105 ns and 6% with τ ) 0.96 ns.
rate constants kET
.
Distance Dependence of Electron-Transfer Rates. The
electron-transfer rate constants kET for the Zn,Fe dimers 12a-f
are presented in Table 2, along with three different measures
of the distances from porphyrin edge to porphyrin edge. For
comparison, the dimers with polyphenylene and bicyclo[3.3.0]-
octylidene spacers are also included. Direct donor-acceptor
coupling is extremely weak at the long distances involved here.
Therefore, the intervening bridge must mediate the donor-
acceptor interaction.
triexponential fit. This is probably a result of multiple confor-
mations, each with its own distance and rate. This dimer is
omitted from subsequent correlations, because its distance is
not well-defined. In contrast, the monoexponential decay of the
other dimers is evidence that their linkers are more rigid and
impose a well-defined spatial relation between the two porphyrin
rings. Even with the steroidal and bicyclooctane linkers, the
possibility of conformational flexibility does not complicate the
decay, so it is unlikely that the more pronounced biexponential
kinetics in 12c and 12f is due to conformational heterogeneity,
nor does the presence of regioisomers of 12e (Table 1, entry
13) lead to discernible deviation from monoexponential kinetics,
inasmuch as the distance between the two porphyrin rings is
It is interesting to compare all of the dimers with the same
13.9-14.4 Å edge-to-edge distance. The rate constant for ET
in the piperazine-linked dimer 12a is 18 times that of the dimer
with a terphenylene spacer.16a Likewise, ET in 12a is faster than
that across the bicyclo[3.3.0]octylidene spacers.19 This difference
is perhaps due to the heterocyclic ring with nitrogen lone pairs
in 12a in place of the alkane, diene, and aromatic rings in the
comparison dimers. However, another possible explanation is
that electron transfer is faster in 12a because it has a low-spin
bis(imidazole)Fe(III) that is already in the porphyrin plane, as

10550 J. Phys. Chem. A, Vol. 103, No. 49, 1999
Portela et al.
decay per ring
TABLE 2: Electron-Transfer Rate Constants and Interporphyrin Distancesfor Dimeric Zn,Fe Porphyrins
structure
bridgea
dspace, Å
dbond, Å
Nbond
τ, ns
0.025
0.15
0.90
10-9 kET, s-1
C6H4
5.8
10.1
14.4
13.9
14.4
14.4
15.2
18.6
20.6
21.4
27.3
7.2
12.9
18.6
5
9
40c
b
b
(C6H4)2
(C6H4)3
C6H4-BCO-C6H4
C6H4-BCOE-C6H4
C6F4-pip-C6F4
C6F4-bcoda-C6F4
C6F4-bip-C6F4
C6F4-stde-C6F4
C6F4-stda-C6F4
C6F4-pip-pfbp-pip-C6F4
6c
0.15
0.13
13
14
14
13
15
17
20
20
25
0.8c
b
d
4.3
8.8
d
12a
12b
12c
12d
12e
12f
18.6
21.7
24.7
29.3
29.4
35.6
0.07 ( 0.005
0.3 ( 0.02
0.36 ( 0.02
0.72 ( 0.02
0.85 ( 0.02
1.40 ( 0.05
14 ( 0.1
3.0 ( 0.01
2.1 ( 0.01
0.75 ( 0.03
0.54 ( 0.02
0.085 ( 0.03
0.15
0.20
a BCO ) bicyclo[3.3.0]octanylidene, BCOE ) bicyclo[3.3.0]octadienylidene; other abbreviations are as in Table 1. b Reference 16a. c Corrected
by subtracting 1/τ0. d Reference 19.
TABLE 3: Comparison of â Values for Electron Transfer
Across Spacer Rings
correl
coeff k(r + 4 Å)
k(r)/
ET system
distance â, Å-1
k0, s-1
polyphenylene dspace
0.45
5.7 × 1011 0.9998
4.8 × 1011 0.9998
1.3 × 1012 0.970
2.0 × 1012 0.988
6
4
spacersa
dbond
0.34
Nbond
0.49b 4.7 × 1011 0.9999
12a-f
dspace
dbond
Nbond
0.36
0.28
4
3
0.40b 1.9 × 1012 0.989
a Reference 16a, corrected with 1/τ0. b Per bond.
clear that eq 4 is followed well regardless of the particular choice
of distance measure.
Figure 3. Distance dependence of ET rates in dimeric Zn,Fe porphyrins
12a-f: dspace (O), dbond (0), Nbond (]), eq 5, with R ) dspace + 6.97 Å
(s).
k(r) ) k0 exp[-â(r-r0)]
(4)
compared to the bicyclo[3.3.0]octylidene dimer, where the Fe-
(III) was high-spin, five-coordinated, and above the porphyrin
plane.
The values of â and k0 obtained from least-squares fitting of
the data in Table 2 for each of the distance measures are
summarized in Table 3. The scatter is greater, and the correlation
coefficient is lower, than that seen with the polyphenylene
series,16a which is also included. However, this latter is
structurally homogeneous, whereas 12a-f include a variety of
intervening atoms. Moreover, k0 near 2 × 1012 is in good
agreement with that seen for electron transfer from a zinc
porphyrin to a directly attached quinone.40 The good linearity
suggests that the current theory of electron tunneling via a
nonconducting spacer ought to be applicable to all of these.
Previous studies of the distance dependence of PET rates in
covalently linked systems have found that â lies between 0.5
and 1.5 Å-1, with the smallest values for bridges involving
mainly π orbitals.9 Yet our value of 0.36 or 0.28 Å-1, depending
on the distance measure chosen, is considerably smaller than
these and slightly smaller than the values for the phenylene
spacers.16a Such low values mean that the ET rates are only
weakly attenuated by increasing the length of the bridge.
Also included in Table 3 is k(r)/k(r + 4 Å), which is the
factor by which the rate is reduced on increasing the distance
by 4 Å and which can be calculated as exp(4â). In previous
cases this reduction factor was 20-fold41 for aromatic spacers
and 100-fold for aliphatic ones.5b The difference has been
attributed to a superexchange mechanism that is more effective
with the aromatic spacers. However, the five-fold reduction
factor for the phenylene spacers, which are also aromatic, is
significantly lower,16a and the reduction factor for dimers 12a-
f, which include both aromatic and aliphatic spacers, is even
lower, only three- or four-fold. These comparisons imply that
the aromatic nature of the spacer is not the only determinant of
the distance dependence.
The long rigid saturated steroid skeleton of the stde and stda
spacers in 12d and 12e represents an insulator between two
tetrafluorophenyl rings. Yet electron transfer through these
dimers is quite fast, not much slower than the other, previously
measured dimers in Table 2, nor is there any deviation associated
with the ether links unique to 10d. Moreover, it is remarkable
that electron transfer in 12f, across the C6F4-pip-pfbp-pip-C6F4
spacer, is fast enough to be measured, albeit with a larger relative
error than that for the shorter spacers. The through-space
distance of 27.3 Å or the through-bond distance of 35.6 Å is
quite long, especially for a saturated spacer, although it is not
as long as the 40 Å across which an electron has been seen to
travel from a guanine to an anthraquinone in duplex DNA,
which is aromatic.37
The data in Table 2 show that there is a consistent decrease
of the ET rate constant with increasing number of bonds,
through-space distance, or through-bond distance. The matrix
element governing ET can be assumed to decrease exponentially
with distance,38 resulting in a distance dependence for the rate
that should follow eq 4, where k0 is the limiting ET rate when
the donor and acceptor are at the van der Waals contact distance
r0 and where â is the decay constant, also called the distance
decay factor. The equation is so simple because matching the
zinc porphyrin with the identical iron porphyrin maintains a
constant energetic driving force.1 Also, this equation incorporates
into â any distance dependence of the solvent reorganization
energy,39 although this is likely to be negligible because the
positive charges are delocalized over porphyrin rings. Figure 3
shows plots of ln kET vs these various distance measures. It is

Electron Transfer in Metalloporphyrin Dimers
J. Phys. Chem. A, Vol. 103, No. 49, 1999 10551
The weak attenuation of ET rates with increasing length of
these aliphatic spacers is surprising. Perhaps electron transfer
through the thick “wire” of the long steroidal bridges, which
have more parallel pathways, competes adequately with the
transfer through the thin “wire” of the aromatic bridges.
However, that does not account for the rate with the C6F4-pip-
pfbp-pip-C6F4 spacer. We conclude that electrons can find a
path to tunnel through the σ bonds of saturated systems and
that nitrogen or oxygen lone pairs facilitate this process.
The distance dependence of ET is regulated by coupling
pathways that are optimal combinations of through-bond,
through-space, and hydrogen-bond links, rather than simply
depending on through-space distance.3 As the tunneling electron
propagates through a covalent bond, its wave function decays
by a factor ꢀ. For proteins an average ꢀ was estimated as 0.6.
Then the wave function decay per ring, with its two paths, is
given by 2ꢀ4, and the rate decay per ring, which is proportional
to the square of the wave function, is 4ꢀ8. The rate decays per
ring for dimers 12c and 12f are included in Table 2, along with
those for the phenylene spacers. From the average of these
decays, a value of 0.67 is obtained, remarkably close to that in
proteins.
Finally, it is necessary to demonstrate that these rates really
do represent electron transfer, rather than energy transfer. Some
experimental results and authorities are cited above,19 including
a transient absorption feature at 600-700 nm in similar Zn,Fe
porphyrin dimers attributable to the Zn porphyrin cation,16a and
that assignment was supported by the influence of solvent
dynamics on electron transfer in cofacial porphyrin dimers.42
Yet the possibility of excitation energy transfer has not been
rigorously excluded. We now estimate the rate of energy transfer
by the Fo¨rster dipole-dipole mechanism, which is given by eq
5,43 where τ0 is the excited-state lifetime in the absence of energy
methods for the preparation for these dimers are presented here.
The rigid bridges allow the investigation of incremental distance
effects on kET for similar collinear orientations, with edge-to-
edge distances ranging from 14.4 to 27.3 Å. The precise control
of the donor-acceptor separation distance allows unambiguous
determination of ET rates from time-resolved fluorescence
measurements. The rates obtained show a remarkably low falloff
of rate with distance, even with the partly saturated linkages.
Acknowledgment. We thank Jose´ N. Onuchic for thoughtful
discussions of bridge-mediated tunneling. This work was
supported by the National Institutes of Health (Research Grant
HL13581) and by Fundac¸a˜o de Amparo a` Cieˆncia e Tecnologia
(Governo do Estado de Pernambuco, Recife, Brazil).
References and Notes
(1) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265.
(2) Sutin, N. In Electron Transfer in Inorganic, Organic, and Biological
Systems; Bolton, J. R., Mataga, N., McLendon, G., Eds.; Advances in
Chemistry Series 228; American Chemical Society: Washington, DC, 1991;
Chapter 3, p 25. Miller, J. R. In Electron Transfer in Inorganic, Organic,
and Biological Systems; Bolton, J. R., Mataga, N., McLendon, G., Eds.;
Advances in Chemistry Series 228; American Chemical Society: Wash-
ington, DC, 1991; Chapter 17, p 265. Gray, H. B.; Winkler, J. R. Annu.
ReV. Biochem. 1996, 65, 537. Langen, R.; Chang, I. J.; Germanas, J. P.;
Richards, J. H.; Winkler, J. R.; Gray, H. B. Science 1995, 268, 1733.
(3) Onuchic, J. N.; Beratan, D. N. J. Am. Chem. Soc. 1987, 109, 6771.
Beratan, D. N.; Onuchic, J. N.; Winkler, J. R.; Gray, H. B. Science 1992,
258, 1740. Onuchic, J. N.; Beratan, D. N.; Winkler, J. R.; Gray, H. B.
Annu. ReV. Biophys. Biomol. Struct. 1992, 21, 349-377. Beratan, D. N. J.
Am. Chem. Soc. 1986, 108, 4321. Beratan, D. N.; Onuchic, J. N.; Betts, J.
N.; Bowler, B. E.; Gray, H. B. J. Am. Chem. Soc. 1990, 112, 7915. Skourtis,
S. S.; Regan, J. J.; Onuchic, J. N. J. Phys. Chem. 1994, 98, 3379.
(4) Turro, N. J.; Barton, J. K. J. Biol. Inorg. Chem. 1998, 3, 201. Lewis,
F. D.; Letsinger, R. L. J. Biol. Inorg. Chem. 1998, 3, 215. Krider, E. S.;
Meade, T. J. J. Biol. Inorg. Chem. 1998, 3, 222.
(5) (a) Joran, A. D.; Leland, B. A.; Felker, P. M.; Zewail, A. H.;
Hopfield, J. J.; Dervan, P. B. Nature 1987, 327, 508. (b) Joran, A. D.;
Leland, B. A.; Geller, G. G.; Hopfield, J. J.; Dervan, P. B. J. Am. Chem.
Soc. 1984, 106, 6090. (c) Leland, B. A.; Joran, A. D.; Felker, P. M.;
Hopfield, J. J.; Zewail, A. H.; Dervan, P. B. J. Phys. Chem. 1985, 89, 5571.
(d) Khundkar, L. R.; Perry, J. W.; Hanson, J. E.; Dervan, P. B. J. Am.
Chem. Soc. 1994, 116, 9700.
(6) McLendon, G. Acc. Chem. Res. 1988, 21, 160. Meyer, T. J. Acc.
Chem. Res. 1989, 22, 163. Gust, D.; Moore, T. A. Science 1989, 244, 35.
Gust, D.; Moore, T. A. Top. Curr. Chem. 1991, 159, 103. Macpherson, A.
N.; et al. J. Am. Chem. Soc. 1995, 117, 7202.
(7) Connolly, J. S.; Bolton, J. R. In Photoinduced Electron Transfer;
Fox, M. A., Chanon, M., Eds.; Elsevier: Amsterdam, The Netherlands,
1988; Chapter 6.2, Part A. Balzani, V.; Scandola, F. Supramolecular
Photochemistry; Ellis Horwood: New York, 1991; pp 110-121 and
references therein. Hayashi, T.; Ogoshi, H. Chem. Soc. ReV. 1997, 26, 355.
(8) Bowler, B. E.; Raphael, A. L.; Gray, H. B. Prog. Inorg. Chem.
1990, 38, 259. Winkler, J. R.; Gray, H. B. Chem. ReV. 1992, 92, 369. Mayo,
S. L.; Ellis, W. R., Jr.; Crutchley, R. J.; Gray, H. B. Science 1986, 233,
948.
(9) Wasielewski, M. R. Chem. ReV. 1992, 92, 435. Sakata, Y.; Tsue,
H.; O’Neil, M. P.; Wiederrecht, G. P.; Wasielewski, M. R. J. Am. Chem.
Soc. 1994, 116, 6904.
(10) Osuka, A.; et al. Bull. Chem. Soc. Jpn. 1992, 65, 2807. Osuka, A.;
Maruyama, K. Pure Appl. Chem. 1990, 62, 1511.
kE ) (1/τ0)(R0/R)6
(5)
transfer, R is the distance between donor and acceptor transition
dipoles, and R0 ) (9000K2Q ln 10/128π5n4N)∫(fD(ν) ꢀA(ν)/ν4)
dν, where K2 is an orientation factor between 0 and 1, Q is the
donor fluorescence quantum yield in the absence of energy
transfer, n is the refractive index of the medium, N is Avogadro’s
number, ν is frequency in cm-1, fD(ν) is the fluorescence
spectrum of the donor normalized to unit area, and ꢀA(ν) is the
decadic molar absorptivity of the acceptor in units of M-1 cm-1
,
leading to R0 ∼ 20 Å. Using τ0 ) 1.6 ns from Table 1 leads to
the continuous line in Figure 3, plotted versus dspace, the
difference between R and twice the distance from the porphyrin
center to its edge (6.97 Å). It is evident that the Fo¨rster
mechanism is not fast enough to explain the experimental data.
Moreover, it does not account for the experimental linearity.
Therefore, we conclude that eq 3 is indeed valid for the rate of
electron transfer, but we cannot exclude a small contribution
from energy transfer.44
(11) Chambron, J. C.; Harriman, A.; Heitz, V.; Sauvage, J. P. J. Am.
Chem. Soc. 1993, 115, 7419. Kurreck, H.; Huber, M. Angew. Chem., Int.
Ed. Engl. 1995, 34, 849.
Further effort will probe the dependence of PET rate on
exergonicity. We are currently employing this sequential build-
ing-block approach to the synthesis of another series of
metalloporphyrin dimers. It is thereby possible to introduce a
variety of different substituents onto the acceptor porphyrin
while maintaining a constant piperazine linker.
(12) Wynne, K.; Lecours, S. M.; Galli, C.; Therien, M. J.; Hochstrasser,
R. M. J. Am. Chem. Soc. 1995, 117, 3749.
(13) Staab, H. A.; Hauck, R.; Popp, B. Eur. J. Org. Chem. 1998, 631.
(14) Osuka, A.; Maruyama, K.; Mataga, N.; Asahi, T.; Yamakazi, I.;
Tamai, N. J. Am. Chem. Soc. 1990, 112, 4958. Osuka, A.; Tanabe, N.;
Kawabata, S.; Yamazaki, I.; Nishimura, Y. J. Org. Chem. 1995, 60, 7177.
(15) Berg, A.; Rachamim, M.; Galili, T.; Levanon, H. J. Phys. Chem.
1996, 100, 8791.
(16) (a) Helms, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc. 1992,
114, 6227. (b) Helms, A.; Heiler, D.; McLendon, G. J. Am. Chem. Soc.
1991, 113, 4325. (c) Heiler, D.; McLendon, G.; Rogalskyj, P. J. Am. Chem.
Soc. 1987, 109, 604.
(17) Bolton, J. R.; et al. In Electron Transfer in Inorganic, Organic,
and Biological Systems; Bolton, J. R., Mataga, N., McLendon, G., Eds.;
Conclusions
A series of metal meso-tetraarylmetalloporphyrin dimers have
been synthesized. These dimers were designed to investigate
critical geometric parameters governing ET rates. Synthetic

10552 J. Phys. Chem. A, Vol. 103, No. 49, 1999
Portela et al.
Advances in Chemistry Series 228; American Chemical Society: Wash-
ington, DC, 1991; Chapter 7, p 118. Newton, M. D. Chem. ReV. 1991, 91,
767.
(18) Sessler, J.; Johnson, M.; Liu, T.; Creager, S. J. J. Am. Soc. Chem.
1988, 110, 3659. Sessler, J. L.; Johnson, M. R.; Creager, S. E.; Fettinger,
J. C.; Ibers, J. A. J. Am. Chem. Soc. 1990, 112, 9310. Sessler, J. L.; Johnson,
M. R.; Lin, T.-Y. Tetrahedron 1989, 45, 4767. Rodriguez, J.; Kirmaier,
C.; Johnson, M. R.; Friesner, R. A.; Holten, D.; Sessler, J. S. J. Am. Chem.
Soc. 1991, 113, 1652.
(19) de Rege, P. J. F.; Williams, S. A.; Therien, M. J. Science 1995,
269, 1409.
(20) Battioni, P.; Brigaud, O.; Desvaux, H.; Traylor, T. G. Tetrahedron
Lett. 1991, 32, 2893.
(30) Osuka, A.; Nakajima, S.; Maruyama, K. J. Org. Chem. 1992, 57,
7355. Osuka, A.; Kobayashi, F.; Nakajima, S.; Maruyama, K.; Yamazaki,
I.; Nishimura, Y. Chem. Lett. 1993, 161. Osuka, A.; Liu, B.-L.; Maruyama,
K. J. Org. Chem. 1993, 58, 3582.
(31) Lindsey, J. S.; Schreiman, I. C.; Hsu, H. S.; Kearney, P. C.;
Marguerettaz, A. M. J. Org. Chem. 1987, 523, 827. Lindsey, J. S.; Wagner,
R. W. J. Org. Chem. 1989, 54, 829.
(32) van der Made, A. W.; Hoppenbrouwer, E. J. H.; Nolte, R. J. M.;
Drenth, W. Recl. TraV. Chim. Pays-Bas 1988, 107, 15.
(33) Longo, F. R.; Finareli, M. G.; Kim, J. B. J. Heterocycl. Chem.
1968, 6, 927.
(34) Kobayashi, H. AdV. Biophys. 1975, 8, 191.
(21) Schmidt, J. A.; Liu, J. Y.; Bolton, J. R.; Archer, M. D.; Gadzekpo,
V. P. Y. J. Chem. Soc., Faraday Trans. 1 1989, 85, 1027.
(22) Magde, D.; Campbell, B. F. SPIE 1989, 1054, 61.
(23) Assis, M. D.; Smith, J. R. L. J. Chem. Soc., Perkin Trans. 2 1998,
2221. Magde, D.; Portela, C. F.; Richards, J. L.; Scho¨llhorn, B.; Traylor,
T. G. 2nd European Bioinorganic Conference on Metal Ions in Biological
Systems (EUROBIC II), Florence, Italy, 1994; Abstract 288.
(24) Smith, C. R. J. Am. Chem. Soc. 1928, 50, 1936.
(25) (a) Della, E. W.; Tsanaktsidis, J. Aust. J. Chem. 1985, 38, 1705.
(b) Adcock, W.; et al. J. Phys. Org. Chem. 1991, 4, 353.
(26) Davis, M.; Parnell, E. W.; Rosenbaum, J. J. Chem. Soc. C 1967,
1045.
(27) Wagner, R. W.; Johnson, T. E.; Li, F. R.; Lindsey, J. S. J. Org.
Chem. 1995, 60, 5266. Lindsey, J. S.; Prathapan, S.; Johnson, T. E.; Wagner,
R. W. Tetrahedron 1994, 50, 8941.
(28) Chang, C. K.; Abdalmuhdi, I. J. Org. Chem. 1983, 48, 5388. Chang,
C. K.; Abdalmuhdi, I. Angew. Chem., Int. Ed. Engl. 1984, 23, 164. Eaton,
S. S.; Eaton, G. R.; Chang, C. K. J. Am. Chem. Soc. 1985, 107, 3177.
Fillers, J. P.; Ravichandran, K. G.; Abdalmuhdi, I.; Tulinsky, A.; Chang,
C. K. J. Am. Chem. Soc. 1986, 108, 417.
(35) Ruzicka, L.; Goldberg, M. W. HelV. Chim. Acta 1936, 19, 99.
(36) Scheidt, W. R.; Reed, C. A. Chem. ReV. 1981, 81, 543. Scheidt,
W. R. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York,
1978; Vol. 3A, p 463. Collins, D. M.; Countryman, R.; Hoard, J. L. J. Am.
Chem. Soc. 1972, 94, 2066.
(37) Gasper, S. M.; Schuster, G. B. J. Am. Chem. Soc. 1997, 119, 12762.
(38) Hopfield, J. J. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 3640.
(39) Isied, S. S.; Vassilian, A.; Wishart, J. F.; Creutz, C.; Schwarz, H.
A.; Sutin, N. J. Am. Chem. Soc. 1988, 110, 635.
(40) Wynne, K.; LeCours, S. M.; Galli, C.; Therien, M. J.; Hochstrasser,
R. M. J. Am. Chem. Soc. 1995, 117, 3749.
(41) Finckh, P.; Heitele, H.; Volk, M.; Michel-Beyerle, M. E. J. Phys.
Chem. 1988, 92, 6584.
(42) Zaleski, J. M.; Chang, C. K.; Nocera, D. G. J. Phys. Chem. 1993,
97, 13206.
(43) Berlman, I. B. Energy Transfer Parameters of Aromatic Com-
pounds; Academic Press: New York, 1973; pp 27 and 28.
(44) Yang, S. I.; Lammi, R. K.; Seth, J.; Riggs, J. A.; Arai, T.; Kim,
D.; Bocian, D. F.; Holten, D.; Lindsey, J. S. J. Phys. Chem. 1998, 102,
9426.
(29) Collman, J. P.; et al. J. Am. Chem. Soc. 1992, 114, 9869. Guilard,
R.; et al. J. Am. Chem. Soc. 1992, 114, 9877.